Re: [b2g] Nightly OTA update on Flame stops?
Marek Raida writes: > Hmm, I've another problem, but similar. OTA is uploaded to phone but > never applied. Is somebody experiencing the same problems?? > > I wrote more to the bug, however no one has commented it yet... > > https://bugzilla.mozilla.org/show_bug.cgi?id=1160879 This had been reported before. It looks like your internal storage has not enough space to extract the whole update. Remove some fonts/music file to free the space up might help. For example: https://bugzilla.mozilla.org/show_bug.cgi?id=1079786 https://bugzilla.mozilla.org/show_bug.cgi?id=1073399 Kanru ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
Re: [b2g] Nightly OTA update on Flame stops?
Looking at the size space in the bug you wrote, it seems similar to bug 1155704 (which we had recently). Looking at your log, I see : I/GonkAutoMounter(15835): Mounted /system partition as read-only Do you know what base build you are on? If you are on linux or mac, you can use this python script : https://github.com/Mozilla-TWQA/B2G-flash-tool/blob/master/check_versions.py or this shell script : https://github.com/nhirata/B2G-flash-tool/blob/master/check_versions.sh to show the gecko/gaia and base build (bootloader) you are on. I suppose I should reply this in the bug. We can carry on the conversation there if you would like. Regards, Naoki On May 6, 2015, at 11:03 PM, Marek Raida wrote: > Hmm, I've another problem, but similar. OTA is uploaded to phone but never > applied. Is somebody experiencing the same problems?? > > I wrote more to the bug, however no one has commented it yet... > > https://bugzilla.mozilla.org/show_bug.cgi?id=1160879 > > Thx a lot for advice, > Marek > ___ > dev-b2g mailing list > dev-b2g@lists.mozilla.org > https://lists.mozilla.org/listinfo/dev-b2g ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
Re: [b2g] Failure: `BRANCH=v1-train ./config.sh emulator`
On 06/05/2015 22:38, Felix E. Klee wrote: > Then I built the `emulator-jb` B2G, and the boot image. > > Attempt to boot the device with that image: > > $ fastboot boot myboot.img > downloading 'boot.img'... > OKAY [ 0.302s] > booting... > FAILED (remote: dtb not found) > finished. total time: 0.367s > > Seems like the DTB, the Device Tree Blob, is missing from the image. > Is that because I'm using the emulator B2G? I guess it is. Try the `flame` build again but do this first: before invoking build.sh delete the file under device/t2m/flame/extract-files.sh This should prevent the script from running and in turn allows you to build the system without pulling the binary blobs. Those are needed only if you need to re-flash the entire /system partition image but you won't be doing that obviously so it should be fine. Gabriele signature.asc Description: OpenPGP digital signature ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
Re: [b2g] Failure: `BRANCH=v1-train ./config.sh emulator`
Le 06/05/2015 22:38, Felix E. Klee a écrit : > Got the kernel to compile. Actually, there were still errors, which I > didn't see due to `-j4`. `zImage` was not build. I had to adjust some > include paths, weird. > > Then I built the `emulator-jb` B2G, and the boot image. > > Attempt to boot the device with that image: > > $ fastboot boot myboot.img > downloading 'boot.img'... > OKAY [ 0.302s] > booting... > FAILED (remote: dtb not found) > finished. total time: 0.367s > > Seems like the DTB, the Device Tree Blob, is missing from the image. > Is that because I'm using the emulator B2G? The DTB must be generated (I think for msm8610 it is done by the kernel build) and then integrated within the boot partition. You can have a look at Z3 Compact port, it makes use of DTB. > ___ > dev-b2g mailing list > dev-b2g@lists.mozilla.org > https://lists.mozilla.org/listinfo/dev-b2g > ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
Re: [b2g] Exposing the browser api reference to the internal/certified apps(eg. System app) themselves?
Tim Guan-tin Chien writes: > We did talk about these solutions and neither is desirable as they > adds significant complexity to the System app internal, and the cost > of reviewing every future patch. > > (We even went to not-so-crazy ideas like simply overwrite window.Audio > with our wrap) > > That's why a platform API is better and uniform IMHO. Alternatively, how about wrapping the audio playing part of system app in a separated browser-element? Then you would control the audio just like controlling other apps. Kanru ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
Re: [b2g] Failure: `BRANCH=v1-train ./config.sh emulator`
On Thu, May 7, 2015 at 10:46 AM, Gabriele Svelto > I guess it is. Try the `flame` build again but do this first: before > invoking build.sh delete the file under > device/t2m/flame/extract-files.sh Thanks, the build now finished: $ cd B2G $ BRANCH=v1.4 ./config.sh -j4 flame $ rm device/t2m/flame/extract-files.sh $ ./build.sh -j4 $ cd .. $ ./core/mkbootimg/mkbootimg \ --kernel OT_4019X/kernel/arch/arm/boot/zImage \ --ramdisk B2G/out/target/product/flame/ramdisk.img \ --base 0x20 --cmdline 'androidboot.hardware=qcom loglevel=1' \ -o myboot.img However, there is still the message about the missing DTB when booting. I may try Alexandre’s suggestion, that is if I can find information about that port. Searching the web for “Firefox OS” and “Z3 compact” wasn’t successful for me. ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
Re: [b2g] Failure: `BRANCH=v1-train ./config.sh emulator`
On Thu, May 7, 2015 at 10:46 AM, Gabriele Svelto > I guess it is. Try the `flame` build again but do this first: before > invoking build.sh delete the file under > device/t2m/flame/extract-files.sh Thanks, the build now finished: $ cd B2G $ BRANCH=v1.4 ./config.sh -j4 flame $ rm device/t2m/flame/extract-files.sh $ ./build.sh -j4 $ cd .. $ ./core/mkbootimg/mkbootimg \ --kernel OT_4019X/kernel/arch/arm/boot/zImage \ --ramdisk B2G/out/target/product/flame/ramdisk.img \ --base 0x20 --cmdline 'androidboot.hardware=qcom loglevel=1' \ -o myboot.img However, there is still the message about the missing DTB when booting. I may try Alexandre’s suggestion, that is if I can find information about that port. Searching the web for “Firefox OS” and “Z3 compact” wasn’t successful for me. ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
[b2g] A talk on p2p, mozTCPSocket & friends
Hey b2g'ers, I recently gave a talk on some of the stuff coming out from b2g at EmpireJS in NY. Reception was really enthusiastic and people were very excited about being able to do things such as running a web server and controlling NFC using simple JS calls, then running web components + web audio based instruments on various phones (Firefox OS, Android). I think they really liked the interoperability aspect too :-) I showed the talk to Justin, who's worked on a few of the materials I am using as a base to build my project, and he said you could maybe be interested in having a look at it--so there it is: https://www.youtube.com/watch?v=xrWH3WO0xrE I'm still working on this project, as what is shown in the talk is a very early prototype, so there are going to be refinements in the future. If you're interested in peeking at the code here is the (temporary, somewhat anonymous) repo for the server app which is running on the Nexus 4: https://github.com/sole/d-e-impl Thanks to everyone whose work I'm building upon: NFC, fxos-web-server, wifiManager, mozTCPSocket... I'm really excited about this whole world we can build with FxOS, and other developers are excited too, which is totally awesome! Keep up the good work \o/ sole -- http://soledadpenades.com ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
Re: [b2g] Failure: `BRANCH=v1-train ./config.sh emulator`
Le 07/05/2015 14:02, Felix E. Klee a écrit : > On Thu, May 7, 2015 at 10:46 AM, Gabriele Svelto >> I guess it is. Try the `flame` build again but do this first: before >> invoking build.sh delete the file under >> device/t2m/flame/extract-files.sh > > Thanks, the build now finished: > > $ cd B2G > $ BRANCH=v1.4 ./config.sh -j4 flame > $ rm device/t2m/flame/extract-files.sh > $ ./build.sh -j4 > $ cd .. > $ ./core/mkbootimg/mkbootimg \ > --kernel OT_4019X/kernel/arch/arm/boot/zImage \ > --ramdisk B2G/out/target/product/flame/ramdisk.img \ > --base 0x20 --cmdline 'androidboot.hardware=qcom loglevel=1' \ > -o myboot.img You should be able to add a --dt flag and pass it the dt.img generated > > However, there is still the message about the missing DTB when booting. > I may try Alexandre’s suggestion, that is if I can find information > about that port. Searching the web for “Firefox OS” and “Z3 compact” > wasn’t successful for me. > ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
Re: [b2g] 3.0 Flame testing, the pretend-256M 319M memory threshold, , , and GC settings
> On 03/25/2015 5:49 PM, Kyle Huey wrote: > > Previous experience suggests that if we relax the limits at all it's going > to be quite difficult to claw that back. > > Somebody should take some memory reports from 2.2 and trunk and diff them, > so we can start looking at where the fat is. After much procrastinating I finally managed to carve some time to do a diff between v2.1, v2.2 and master and frankly the results are worrisome. I've made a clean install of the three engineering builds and opened some applications with empty workloads (dialer, sms, gallery, music, clock and settings) and gathered memory reports after having minimized all of them. Here's the bird's eye view of the three branches by USS: NAME v2.1 v2.2 master b2g 53.3 61.2 64.6 (Nuwa)2.35.58.9 Homescreen 10.6 13.8 24.1 Built-in Keyboa 11.6 12.6 13.7 Communications9.9 10.8 16.7 Messages 15.4 17.1 17.3 Gallery8.78.6 14.9 Music 13.79.0 15.4 Clock9.0 14.6 15.4 Settings 17.4 13.9 20.5 (Preallocated)5.15.2 - Yep, it's that bad. The main process gained a full 13.3 MiB of USS, that used to be enough to run an entire app only some time ago. Practically all applications are well above 10MiB at startup (they almost all used to be below that) and many of them take almost twice as much as they used to. The preallocated process is gone (absorbed into Nuwa IIRC) but even then the aggregated result is 1.1MiB larger. The regressions in master are *huge* but v2.2 was already in very poor shape (and I haven't even tried with v2.0). I did not even check by populating with the reference workloads mostly because of lack of time but I don't think that's going to look pretty either. I'm now looking at the about:memory reports to figure out where this is coming from. Do we already have a bug about memory regressions on v3.0/master? If not I'll open one and attach everything there. Gabriele signature.asc Description: OpenPGP digital signature ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
Re: [b2g] 3.0 Flame testing, the pretend-256M 319M memory threshold, , , and GC settings
On 05/07/2015 02:59 PM, Gabriele Svelto wrote: On 03/25/2015 5:49 PM, Kyle Huey wrote: Previous experience suggests that if we relax the limits at all it's going to be quite difficult to claw that back. Somebody should take some memory reports from 2.2 and trunk and diff them, so we can start looking at where the fat is. After much procrastinating I finally managed to carve some time to do a diff between v2.1, v2.2 and master and frankly the results are worrisome. I've made a clean install of the three engineering builds and opened some applications with empty workloads (dialer, sms, gallery, music, clock and settings) and gathered memory reports after having minimized all of them. Here's the bird's eye view of the three branches by USS: Is there anybody looking into adding this procedure (or similar) as performance tests which are reported daily? In order to know when this issue is fixed, and to watch for regressions later. -- Nicolas B. Pierron ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
Re: [b2g] 3.0 Flame testing, the pretend-256M 319M memory threshold, , , and GC settings
>From what I understand, launch times and memory usage are both underground >large regressions right now from the preallocated process bug: https://bugzilla.mozilla.org/show_bug.cgi?id=1151672 Also, Raptor tracks memory usage for applications as part of the cold launch test, and is run with incoming Gaia commits: http://raptor-ui.divshot.io/#/dashboard/script/memory.js?device=flame-kk&branch=master&memory=319&suite=coldlaunch Thanks, Eli Perelman On Thu, May 7, 2015 at 8:32 AM, Nicolas B. Pierron wrote: > On 05/07/2015 02:59 PM, Gabriele Svelto wrote: >>> On 03/25/2015 5:49 PM, Kyle Huey wrote: >>> >>> Previous experience suggests that if we relax the limits at all it's going >>> to be quite difficult to claw that back. >>> >>> Somebody should take some memory reports from 2.2 and trunk and diff them, >>> so we can start looking at where the fat is. >> >> After much procrastinating I finally managed to carve some time to do a >> diff between v2.1, v2.2 and master and frankly the results are >> worrisome. I've made a clean install of the three engineering builds and >> opened some applications with empty workloads (dialer, sms, gallery, >> music, clock and settings) and gathered memory reports after having >> minimized all of them. Here's the bird's eye view of the three branches >> by USS: > Is there anybody looking into adding this procedure (or similar) as > performance tests which are reported daily? In order to know when this > issue is fixed, and to watch for regressions later. > -- > Nicolas B. Pierron > ___ > dev-b2g mailing list > dev-b2g@lists.mozilla.org > https://lists.mozilla.org/listinfo/dev-b2g___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
Re: [b2g] A talk on p2p, mozTCPSocket & friends
This is really cool - especially NFC cats ;-) Chris Mills Senior tech writer || Mozilla developer.mozilla.org || MDN cmi...@mozilla.com || @chrisdavidmills > On 7 May 2015, at 13:28, Soledad Penadés wrote: > > Hey b2g'ers, > > I recently gave a talk on some of the stuff coming out from b2g at EmpireJS > in NY. > > Reception was really enthusiastic and people were very excited about being > able to do things such as running a web server and controlling NFC using > simple JS calls, then running web components + web audio based instruments on > various phones (Firefox OS, Android). I think they really liked the > interoperability aspect too :-) > > I showed the talk to Justin, who's worked on a few of the materials I am > using as a base to build my project, and he said you could maybe be > interested in having a look at it--so there it is: > https://www.youtube.com/watch?v=xrWH3WO0xrE > > I'm still working on this project, as what is shown in the talk is a very > early prototype, so there are going to be refinements in the future. > > If you're interested in peeking at the code here is the (temporary, somewhat > anonymous) repo for the server app which is running on the Nexus 4: > https://github.com/sole/d-e-impl > > Thanks to everyone whose work I'm building upon: NFC, fxos-web-server, > wifiManager, mozTCPSocket... I'm really excited about this whole world we can > build with FxOS, and other developers are excited too, which is totally > awesome! > > Keep up the good work \o/ > > sole > > -- > http://soledadpenades.com > > ___ > dev-b2g mailing list > dev-b2g@lists.mozilla.org > https://lists.mozilla.org/listinfo/dev-b2g ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
Re: [b2g] 3.0 Flame testing, the pretend-256M 319M memory threshold, , , and GC settings
On 07/05/2015 15:48, Eli Perelman wrote: > From what I understand, launch times and memory usage are both > underground large regressions right now from the preallocated process bug: > > https://bugzilla.mozilla.org/show_bug.cgi?id=1151672 Thanks for pointing that out. I've glanced over the about:memory reports and while some regressions appear genuine (e.g. the system app) in other apps the increase in USS is not matched by an equal increase in explicit allocations. This points out to large chunks of memory which were previously shared with the originating process are potentially not shared anymore. Gabriele signature.asc Description: OpenPGP digital signature ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
Re: [b2g] 3.0 Flame testing, the pretend-256M 319M memory threshold, , , and GC settings
I've filed bug 1151672 [1] with full memory reports attached. I've CC'd some people who might be interested but feel free to add more. It seems that bug 1151672 [2] is responsible for most of the USS increase but I've spotted some other issues outside of that, some of which look like genuine regressions. Gabriele [1] Major memory usage regressions in v3.0 https://bugzilla.mozilla.org/show_bug.cgi?id=1162535 [2] Preallocated process approach not working on 2.2 and 3.0 https://bugzilla.mozilla.org/show_bug.cgi?id=1151672 signature.asc Description: OpenPGP digital signature ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
Re: [b2g] Failure: `BRANCH=v1-train ./config.sh emulator`
On Thu, May 7, 2015 at 1:28 PM, Alexandre Lissy wrote: > You should be able to add a --dt flag and pass it the dt.img generated `dt.img` is missing, and so I built it myself, and added it: $ B2G/device/qcom/common/dtbtool/dtbtool -o dt.img -s 2048 \ -p B2G/kernel/scripts/dtc OT_4019X/kernel/arch/arm/boot $ B2G/system/core/mkbootimg/mkbootimg \ --kernel OT_4019X/kernel/arch/arm/boot/zImage \ --ramdisk B2G/out/target/product/flame/ramdisk.img \ --base 0x20 --cmdline 'androidboot.hardware=qcom loglevel=1' \ --dt OT_4019X/kernel/arch/arm/boot/msm8610-sim.dtb -o myboot.img Still the same error message when attempting to boot: FAILED (remote: dtb not found) Could it be that deleting `device/t2m/flame/extract-files.sh` prevents a proper `dt.img` from being built? ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g
Re: [b2g] Problem flashing gaia - ZTE Open C
hello, Im just tried the other day to install the android kitkat to my Open C and it was ok, but I need to root it and unlocked it, can you show me any link? please? ___ dev-b2g mailing list dev-b2g@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-b2g